Model Switcher
PassAudited by ClawScan on May 1, 2026.
Overview
This instruction-only skill coherently performs model switching, but users should notice that it can automatically change which model handles a task.
This skill appears safe for its stated purpose. Before installing, check that the model names and aliases match your own trusted OpenClaw setup and that you are comfortable with the agent switching models automatically during a conversation.
Findings (2)
Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.
The agent may switch models during a conversation, which can affect response quality, latency, cost, and consistency.
The skill gives the agent instructions to change the active model automatically. This is central to the stated purpose, but it is still a behavior-control capability users should notice.
Use `session_status` tool with `model` parameter: session_status({ model: "kiro-cli" }) ... session_status({ model: "haiku" })Install only if you want automatic model switching; verify that the configured model aliases are the ones you expect, or modify the workflow to ask before switching.
Complex or simple tasks may be sent to the named configured model routes instead of the current/default model.
The skill references specific custom model routes. Since model routing determines where prompts are processed, users should confirm these aliases are part of their trusted OpenClaw configuration.
**Fast model (haiku):** `custom-kiro-cli-vipdump-eu-org/claude-haiku-4-5` ... **Powerful model (sonnet):** `custom-kiro-cli-vipdump-eu-org/claude-sonnet-4-5`
Confirm the listed model aliases and provider routes are trusted and appropriate for the data you plan to share with the agent.